The cost of bringing a drug to market is astounding and the failure rate is daunting. The limited success of conventional drug discovery is in a large part attributed to the wide adoption of a reductionist model of one- drug-one-gene-one-disease. New methodologies are very much called for: Polypharmacology focuses on defining multiple targets to a single drug and studying the effect of these drugs on perturbing disease-causing networks. Drug repurposing reuses existing drugs for new clinical indications. These two modalities have emerged as new drug discovery paradigms, and are strongly prompted by the NIH. However, rational and effective polypharmacology and drug repurposing is currently hindered by our limited understanding of structural and energetic origins of genome-wide drug-target interactions. To address this challenge, this proposal seeks to develop and experimentally validate an innovative methodology to determine high-resolution drug-target interactions on a genome scale. Building on our successful proof-of-concept studies, and close multidisciplinary collaborations between experimental and computational laboratories, we will integrate big data from chemical, structural, and functional genomics, and synthesize techniques derived from large-scale graph mining, global set statistics, chemoinformatics, bioinformatics, molecular modeling, and biophysics. Specifically, we will develop a new chemical similarity search method, ligand Enrichment of Network Topological Similarity (ligENTS), to map the continuous chemical universe to its global pharmacological space. We will integrate ligENTS with our already successful structural systems biology platform to construct high- resolution drug-target interaction models across species and across fold space. To demonstrate the feasibility and innovation of our proposed integrative approach, we will apply it to a test case: anti-infective drug repurposing. The emergence of drug resistant microbes to antibiotics poses a great threat to human health. Repurposing safe drugs to target pathogen-associated proteins has emerged as a novel concept to combat drug resistant pathogens. However, significant technical barriers exist in applying existing drug repurposing strategies across species in the context of pathogen-host interactions. Our innovative approach consolidates chemical, structural and network views of molecular components and their interactions in a biological system, thereby providing a new solution to discovering safe and efficient anti-infective agents by determining molecular targets of bioactive compounds in both humans and pathogens. We will experimentally validate novel pathogen-associated proteins and anti-infective drugs, generated from our in silico predictions, both in vitro and in vivo. If successful, this work will provide the scientific community and pharmaceutical industry with: (a) fundamentally new algorithms and associated software for identifying three-dimensional drug-target interaction models on a genome scale, and (b) experimentally validated novel anti-infective compounds and targets, with the potential to combat antibiotic resistance.